TESTIMONY OF RAND E. ROSENBLATT
PROFESSOR OF LAW & ASSOCIATE DEAN FOR ACADEMIC AFFAIRS
RUTGERS UNIVERSITY LAW SCHOOL
CAMDEN, NEW JERSEY
BEFORE THE SUBCOMMITTEE ON EMPLOYER-EMPLOYEE RELATIONS
OF THE COMMITTEE ON EDUCATION AND THE WORKFORCE OF THE
UNITED STATES HOUSE OF REPRESENTATIVES
HEARING ON "ERISA: A QUARTER CENTURY OF PROVIDING WORKERS HEALTH INSURANCE"
WEDNESDAY, FEBRUARY 24, 1999
2175 RAYBURN HOUSE OFFICE BUILDING
Professor Rand E. Rosenblatt
Rutgers University Law School
217 N. Fifth St.
Camden, NJ 08102
tel: (609) 225-6379
FAX: (609) 225-6516
Chairman Boehner and members of the Committee: Thank you very much for the opportunity to testify on the vital issue of the Employment Retirement Income Security Act (ERISA)’s impact on the more than 120 million Americans whose group health insurance is based on private employment. As a professor of health law, and the lead author of Law and the American Health Care System, I have been writing and teaching about this matter for more than 20 years, to law students and also to law professors, attorneys, physicians, health policy analysts, and federal and state judges, and I am pleased and honored to be able to offer my assistance in this forum. Because of the very limited time available to me to prepare for this hearing, this written testimony focuses on only a few of the most important issues. I request that I be granted the opportunity to supplement my written and oral testimony with additional written comments for the record.
The witness invitation letter describes this hearing as focusing on "the importance of [ERISA], and the preemption of state law that it affords, in providing health insurance to millions of Americans." The background assumption of this hearing appears to be that ERISA and its preemption of state law was intended to be, and has functioned as, a positive force in "providing health insurance to millions of Americans." The argument has been made, and will undoubtedly be made today, that by "freeing" private employment health plans and, to a significant extent, their health insurance issuers from costly state regulation and liability, the federal ERISA statute has enabled "the market" to provide American workers with broad health benefits at a reasonable price.
Although I realize that this view is prominently espoused by many in American business, I must respectfully disagree. First, ERISA preemption was enacted in 1974 not as a fairly-debated policy choice to promote market competition in fringe benefits, but rather through a major failure of the democratic process. Second, far from fostering stable and/or expanding health insurance coverage, ERISA preemption has contributed to the steady erosion of health insurance coverage from the 1980s to the present. Indeed, as interpreted and implemented in the 1980s and 1990s, ERISA preemption has undermined basic principles of insurance, de-stabilized coverage and decisions about coverage, permitted reckless interference with quality of care, violated fundamental principles of accountability in American law, created larger numbers of uninsured (in part by hobbling state-level reforms), and undermined responsible competition itself. The coverage mandates and regulatory and liability provisions under consideration in this Congress, and so vociferously opposed by employers and health insurers, are responses to these problems and an attempt to create more defensible conditions of competition in health insurance and health care delivery.
ERISA Preemption and the Failure of the Democratic Process
The overwhelming purpose of the ERISA statute as enacted in 1974 was not to create "free markets" in employment fringe benefits, but rather to regulate the fringe benefit market in order to respond to two types of massive market failure. The first was the frequent refusal of management and labor to set aside actuarially adequate assets to fund collectively-bargained pensions. When companies and whole industries went into economic decline in the 1960s, they were able to pay only a small fraction or none of the pensions they had promised to long-term employees. Second, pre-ERISA state and federal law did not provide adequate remedies for malfeasance by the trustees of fringe benefit funds. The congressional hearings leading up to ERISA were filled with accounts of trustees engaged in self-dealing and otherwise diverting funds from the intended beneficiaries. ERISA responded to these failures of the employee benefit market by regulating and insuring pension plan assets, and by imposing federal fiduciary duties and legal remedies on benefit plan decisionmakers. See Rand E. Rosenblatt, Sylvia A. Law, and Sara Rosenbaum, Law and the American Health Care System 159-161 (1997), and sources cited therein.
Although Congress had exhaustively considered the bills leading up to ERISA for many years, the law’s final version of preemption was the product of last-minute changes in the conference committee that were never subject to reasonable scrutiny by the democratic process. Prior to the conference committee, both houses of Congress had passed bills that, in the usual way of federal statutes, preempted only state law regarding "subject matters regulated by this Act." Moreover, the bills had always contained the insurance saving clause, inserted at the request of the National Association of Insurance Commissioners and consistent with traditional federal deference to state insurance regulation expressed in the McCarran-Ferguson Act, that nothing in the Act "shall be construed to exempt or relieve any person from any law of any State which regulates insurance." ERISA § 514(b)(2)(A), 29 U.S.C. § 1144(b)(2)(A). However, ten days before final congressional action on ERISA, the House-Senate conference committee proposed two major changes. First, the preemption provision was broadened to cover "any and all State laws insofar as they may now or hereafter relate to any employee benefit plan * * * ", ERISA § 514(a), 29 U.S.C. § 1144(a), including, arguably, matters not regulated by the Act. Second, although the clause saving state insurance law still remained in the Act, it was now qualified by what came to be called the "deemer clause," stating that "no employer benefit plan * * * shall be deemed an insurance company" for purposes of state insurance law. ERISA § 514(b)(2)(B), 29 U.S.C. § 1144(b)(2)(B).
These momentous changes were passed by the full Congress without serious discussion of their significance. In the conference committee report and in statements on the House and Senate floor, the conferees focused on the benefits of uniform national standards and the need to avoid potentially diverse and conflicting state and local requirements that would impinge on the benefit plans of companies that operated in many different states. See Metropolitan Life Insurance Co. v. Massachusetts, 471 U.S. 724, 745-46 nn.22-24 (1985)(setting forth some of the preemption provision's legislative history); Fort Halifax Packing Co. v. Coyne, 482 U.S. 1, 9 (1987). The conferees did not address, much less explain, how uniformity would be achieved by exempting state insurance laws from preemption, and then exempting ERISA plans themselves (as distinct from their insurers) from state insurance regulation. Nor did the conferees address why a century of state insurance law and doctrine should be potentially abandoned with no substantive federal standards being put in their place.
Daniel Fox and Daniel Schaffer's excellent study of the legislative history reveals that the expanded preemption provisions were the product of powerful but narrowly-focused interest groups. Although there is no reference to these events in the legislative history, based on interviews in the mid-1980s with key participants in the 1974 events, Fox and Schaffer conclude that the motivation behind the last-minute expansion of ERISA preemption was the desire of certain large corporations and unions who operated multi-state, joint management-labor self-insured health and pension plans to avoid state regulation and taxation of their funds as "insurance," and to increase their freedom of collective bargaining. See Daniel M. Fox & Daniel C. Schaffer, Health Policy and ERISA: Interest Groups and Semipreemption, 14 J. Health Pol., Pol'y & L. 239, 242 & n.13, 243 (1989).
Whatever the merits of the large employers' and unions' desire for uniformity of national contracts, exemption from state insurance taxes, and greater freedom of collective bargaining, the ERISA preemption provision as enacted was vastly broader than these concerns. The text of the ERISA preemption provision was not limited to state insurance taxes, collectively-bargained national contracts, multi-state corporations, or even collective bargaining. Rather, it came to be interpreted as displacing virtually all state law regarding any employment-based health plan (except insurance policies themselves), for all private employees, only a small minority of whom (about 10.4 percent) are now represented by unions. The result has been a major shift of power from the traditional "bargaining agents" for most employees regarding health insurance--the state legislatures and insurance departments--to employers and health insurance companies.
According to Fox and Schaffer, this result was not perceived or intended. Key interest groups--including state governments, federal health officials, and the health insurance industry--were not consulted or involved in the final broadening of the preemption section. "[C]ongressional staff and a few lobbyists made a major decision about employee benefits policy--mainly, it turned out, about health insurance--as if it were a technical issue." Fox & Schaffer, 14 J. Health Pol., Pol'y & L. at 244. Indeed, the full significance of this "technical issue" was not apparent because managed care and market competition in health care did not effectively exist in 1974, and because the comparison of state-law tort damages with ERISA contract benefit remedies did not loom large in the pension context, where almost all of Congress’ attention was focused. In short, a momentous policy decision, and perhaps the most sweeping federal preemption provision in American history, were enacted without hearings, consultation with the major stakeholders, public awareness, or even understanding by most (or perhaps all) of Congress. For these reasons, it is fair to say that ERISA preemption was enacted in 1974 through a major failure of the democratic process.
The Triple Impact of ERISA Preemption
As enacted in 1974 and later interpreted by the courts, ERISA preemption has functioned in three major ways. First, and most obviously, the deemer clause noted above has been interpreted to mean that "self-insured" or "self-funded" employment-based health plans are not subject to state insurance (or any other kind of) regulation.1 See Metropolitan Life Insurance Co. v. Massachusetts, 471 U.S. 724 (1985), but see New York State Conference of Blue Cross & Blue Shield Plans v. Travelers Insurance Co., 514 U.S. 645 (1995) and Travelers Insurance Co. v. Pataki, 63 F.3d 89 (2d Cir. 1995)(holding that state-imposed surcharges on hospital bills do not "relate to" any kind of ERISA plan, and hence must be paid by self-insured as well as fully-insured employment-based health plans). Second, some lower federal courts have interpreted the insurance savings clause narrowly, and ERISA § 514's "relate to" language broadly, so as to preempt state laws that attempt or purport to regulate health insurers themselves. See, e.g., Corporate Health Insurance, Inc. v. The Texas Dept. of Insurance, 12 F. Supp. 2d 597 (S.D. Tex. 1998) (preempting, inter alia, that part of 1997 Texas statute prohibiting a health insurance carrier, health maintenance organization, or managed care entity from removing from or refusing to renew a health care provider with its plan "for advocating on behalf of an enrollee for appropriate and medically necessary health care for the enrollee," Texas Civ. Prac. & Rem. § 88.002(f)).
Third, and most dramatically, ERISA’s preemption provision, ERISA § 514(a), and its civil enforcement provision, ERISA § 502(a), have been interpreted by the Supreme Court as preempting lawsuits based on state tort, contract, and other sources of law against insurers for improper processing of claims under an employee benefit plan, and leaving beneficiaries only with the federal remedies provided under ERISA itself. See Pilot Life Insurance Co. v. Dedeaux, 481 U.S. 41 (1987). Since ERISA’s remedial provisions have been repeatedly interpreted as not including compensatory or punitive damages, plan beneficiaries can recover only the health care or other benefits promised under the plan--an irrelevant remedy for someone who has already died or been seriously injured by alleged plan negligence or other malfeasance. See, e.g., Turner v. Fallon Community Health Plan, Inc., 127 F.3d 196 (1st Cir. 1997) and cases cited therein. Although the issue has never been explicitly decided by the Supreme Court, numerous lower federal courts have held or implied that ERISA preempts state-law actions against health insurers and HMOs for damages allegedly caused by negligently applied or otherwise faulty utilization review, physician financial incentives and the nondisclosure thereof, and policies and decisions regarding health benefits coverage. See, e.g., Corcoran v. United HealthCare, 965 F.2d 1321 (5th Cir. 1992); Dukes v. U. S. Healthcare, Inc., 57 F.3d 350 (3d Cir. 1995); Shea v. Esensten, 107 F.3d 625 (8th Cir. 1997); Lancaster v. Kaiser Foundation Health Plan, 958 F. Supp. 1137 (E.D. Va. 1997); but see Pappas v. Asbel, 1998 Pa. LEXIS 2718 (Pa. Supreme Court, Dec. 23, 1998)(holding that under the U.S. Supreme Court’s decisions in Travelers and progeny, "[state law] negligence claims against a health maintenance organization do not `relate to’ an ERISA plan" and hence are not preempted).
ERISA Preemption and the Erosion of Health Insurance Based on Private Employment
The percentage of the American population with private health insurance rose steadily and sharply from 9.1% in 1940 to 81.1% in 1980, but then declined to 71.6% by 1990. See Table prepared by Randall Bovbjerg et al., in Rosenblatt et al., Law and the American Health Care System p.12. This trend continued in the 1990s, with the number of uninsured rising from 14.2% in 1995 to 16.1% in 1997, when 43.4 million people were uninsured. Moreover, a Census Bureau study conducted from 1993 to 1995 indicated that 71.5 million people lacked health insurance for at least part of the year. Strikingly, the low unemployment rates of the late 1990s have not offset these trends; "[t]oday, the vast majority of uninsured persons are employed." Similarly, the percentage of nonelderly employees who received health insurance from their employers dropped from 69.2% in 1987 to 64.2% in 1997. See Robert Kuttner, "The American Health Care System," The New England J. of Medicine 340:163, 164, 248 (January 14 & 21, 1999). As Kuttner summarizes,
The most prominent feature of American health insurance coverage is its slow erosion . . . . The most important trend is the deterioration of employer-provided coverage . . . .Most employers have narrowed the choice of plans and shifted costs to employees by capping the employer’s contribution, choosing plans with higher out-of-pocket premiums, or both. These changes, in turn, have caused some employees to forego coverage for themselves and their families and have also led to underinsurance, since many employees, especially those who receive low wages, cannot afford the out-of-pocket charges. The following trends are also eroding insurance coverage of all types: rising premium costs [both employment-based and in the individual market]. . . .[;] the trend toward temporary and part-time work [which included about 29% of working Americans in 1997][;] a reduction in explicit coverage, most notably pharmaceutical benefits . . . .[;] greater de facto limitations on covered care, especially by . . . HMOs. More stringent utilization reviews and economic disincentives for physicians and hospitals are resulting in denial of care and shifting of costs to patients. . . .
Kuttner, supra, at 163. While many of these trends are wholly or partly independent of ERISA, ERISA preemption heightens their impact in a number of ways, e.g. by preventing the states from attempting reforms to broaden or stabilize coverage, and by enabling employers to reduce explicit and de facto coverage. Particularly disturbing examples are provided by the following cases.
In December, 1987, John McGann began filing claims for reimbursement with his employer’s health insurer for AIDS-related medical treatment. At that point, his employment-based health insurance provided for $1 million lifetime coverage. At least partly in response to their knowledge of McGann’s illness, which at that point was shared by no other employee, his employer and its plan administrator re-designed the plan to be self-insured, increased employee contributions and deductibles, adopted a preferred provider plan, eliminated coverage for chemical dependency treatment, and limited benefits for AIDS-related claims to a lifetime maximum of $5,000, while imposing no special limitation on any other catastrophic illness. Not surprisingly, McGann soon exhausted his reduced lifetime maximum, and filed suit against his employer, alleging that he had been discriminated against because he had exercised his rights under the health plan, in violation of ERISA § 510. The Fifth Circuit held that no discrimination had occurred, because ERISA § 510 only protected employees rights to benefits under the plan, and because the plan had been changed in a procedurally correct fashion, McGann no longer had any rights under the plan to coverage for AIDS-related care. McGann v. H & H Music Co., 946 F.2d 401 (5th Cir. 1991), cert. denied, 506 U.S. 981 (1992). The court noted that "there is a world of difference between administering a welfare plan in accordance with its terms and deciding what those terms are to be. A company acts as a fiduciary in performing the first task, but not the second." Indeed, the Supreme Court has stated that "[e]mployers or other plan sponsors are generally free under ERISA, for any reason at any time, to adopt, modify, or terminate welfare plans. . . . Nor does ERISA establish any minimum participation, vesting, or funding requirements for welfare plans as it does for pension plans. . . . " Curtiss-Wright Corp. v. Schoonejongen, 115 S. Ct. 1223, 1228 (1995).
Not surprisingly, the McGann case generated passionate criticism. In a letter to then-Solicitor General Kenneth Starr, the American Association of Retired Persons, the American Hospital Association, the American Medical Association, the National Commission on AIDS, the National Governors' Association, and the United States Conference of Mayors characterized the Fifth Circuit's decision as an "outrage" that "will add to the ranks of the uninsured." See Kathlynn L. Butler, "Comment: Securing Employee Health Benefits Through ERISA and the ADA," 42 Emory L.J. 1197, 1230 & n.215 (1993). Indeed, in the wake of the McGann decision, many health plans did reduce or eliminate AIDS coverage. Milt Freudenheim, Patients Cite Bias in AIDS Coverage by Health Plans, N.Y. Times, June 1, 1993, at A1, D2. See also Owens v. Storehouse, Inc., 984 F.2d 394 (11th Cir. 1993)(following the same analysis and reaching the same result as McGann, and describing insurance caps of $25,000 on coverage for AIDS, mental illness, and substance abuse). Congressman William Hughes stated that "Workers . . . have discovered they are often protected only until they get sick . . . . [Because of federal court decisions], [t]oday's workers . . . actually have fewer rights to their benefits than they had prior to the enactment of ERISA." 138 Cong. Rec. E3049 (daily ed. Oct. 5, 1992), quoted in Butler, supra, 42 Emory L.J. at 1197.
Some witnesses here today may attempt to defend what happened in McGann: the individual with the expensive illness, the argument may go, must be placed in his own "actuarially-appropriate risk pool" so as to preserve reasonably-priced coverage for the majority of employees who have "standard risks." Cf. Karen A. Clifford & Russel P. Iuculano, Commentary: AIDS and Insurance: The Rationale for AIDS-related Testing, 100 Harv. L. Rev. 1806, 1807, 1808, 1810 (1987). But there was no allegation that McGann’s infection with AIDS had been a "preexisting condition," i.e. that he had somehow "unfairly" slipped into the standard risk pool where he did not in fact belong. Indeed, if insurance companies are justified in "weeding out" high risk individuals before they gain entrance to the standard risk pool, does not fairness require that "standard risk" individuals who gain entrance to the insured pool not be "weeded out" after their risk has materialized? Is not the whole point of the insurance bargain that a "standard risk" individual is insured if the (unlikely) risk does occur? Underwriting and market theory, as usually presented, do not address the question of the length of time of the insurance contract, or that insurers typically reserve the right to "amend the policy at any time," or that employers can terminate one contract and begin another, thereby knowingly "de-selecting" the risks that have materialized. Is there some middle ground between "lifetime vesting" of benefits once illness has been diagnosed, and plenary discretion to cut off or limit benefits to any category of insureds at any time? For an effort to articulate such a middle ground, see Congressman William Hughes' proposed Group Health Plan Nondiscrimination Act of 1993, excerpted and discussed in Butler, supra, 42 Emory L.J. at 1231-1235 & nn.217-233 (legislation would prohibit reduction in benefits on the basis of particular diseases or medical conditions for insureds who are undergoing treatment for such conditions and who have filed claims based on such treatment; differential lifetime benefits could not be based on diseases or conditions unless the plan was the product of collective bargaining or the plan could prove that such differentials were needed to enable the plan to continue). See also EEOC Interim Guidance on Application of ADA to Health Insurance,
1993 Daily Lab. Rep. (BNA) 109 d22 (June 9, 1993), reprinted in Rosenblatt et al., Law and the American Health Care System at 343 (requiring employers to justify on actuarial or business necessity grounds "disability-based distinctions" in health benefits, i.e. those that single out a particular disability, a discrete group of disabilities, or disability in general; see also mixed judicial reaction to the EEOC’s position).
Aside from the arguably unfair, retroactive quality of what happened to McGann, the ability of self-insured health plans under ERISA to deny benefits for any type of condition or service contributes significantly to the erosion of health insurance coverage and adverse selection. While employers and insurance industry leaders often castigate "mandates" as expensive frills of no interest to the average employee, see, e.g., the comments of Richard L. Huber, CEO of Aetna, Inc., in Harvard Magazine, March/April 1999, at 42, 47, many mandates are in fact essential to the affordability of very important services. Without mandates, certain disfavored but important services, such as mental health, cutting-edge cancer therapies, and substance abuse treatment may be dropped by generally healthy people, creating adverse selection (i.e. disproportionately high rate of use) among the people seeking the coverage. This in turn raises the price of that coverage, leading more relatively healthy people to drop it, and further raising the price, putting it further out of reach of the people who need it. The results can include foregone early care, more serious illness, high hospitalization costs imposed on state and local government, and unnecessary disability and death. See, e.g., Metropolitan Life Insurance Co. v. Massachusetts, 471 U.S. 724 (1985)(discussing these effects as justifications for state-law mental health mandates); Hilliard v. BellSouth Medical Assistance Plan, 918 F. Supp. 1016 (S.D. Miss. 1995), and subsequent notes in Rosenblatt et al., Law and the American Health Care System at 277-78 (discussing these effects in connection with a company’s refusal to spread the costs of certain transplant procedures over the whole employee group); comments of Harvard Professor Joseph P. Newhouse in response to Mr. Huber’s comments noted above, that for a number of services, "either everybody has it in their policy or nobody has it," and acknowledging that the mandate that everybody has it may raise the overall price and cause some people to drop the coverage, Harvard Magazine, March/April 1999, at 48.
While the McGann case involved an explicit reduction of coverage, coverage is also widely reduced by "back-door" or less explicit means. Many of these methods may violate the terms of the health plan or otherwise be illegal, but because ERISA preempts state damage remedies and provides no significant retroactive remedies of its own, insurers may have little incentive even to comply with their own plans. Thus in 1994, the utilization review provider (Greenspring) of the Travelers Insurance Co. refused to authorize admission to a private hospital for a suicidal alcoholic, despite the fact that enrollment in a 30-day inpatient detoxification program was a defined benefit of the insurance policy. After six harrowing weeks without effective treatment, the patient managed to kill himself. Because ERISA preempted his wife’s subsequent state-based lawsuit, there was never any trial on the merits of the reasons or justifications for the refusal of coverage. See Andrews-Clarke v. Travelers Insurance Co., 984 F. Supp. 99 (D. Mass. 1997); see also Bedrick v. Travelers Insurance Company, 93 F.3d 149 (4th Cir. 1996)(physical therapy coverage for infant with severe cerebral palsy and spastic quadriplegia reduced from 104 sessions to 15 sessions per year by a physician utilization reviewer with no relevant training or experience on the basis of a criterion--"significant progress"--that was not stated in the plan, health insurance contract or policy, or Travelers' internal guidelines; when later questioned in a deposition about whether the originally-prescribed physical therapy was "medically necessary" or "appropriate"--the criteria actually used in the health insurance contract or policy--the physician making the coverage decision answered that these criteria were not relevant to her decision).